[EXP] perf(decode): approx. 100x speed improvement w/ various optimizations

[sfyll]

Motivation

Faced with millions of transactions, Heimdall wasn't nearly fast enough, decoding calldata at a rate of approx. 10,000 txs per min, with a cap at length 15,000 bytes (I know gas or gas-normalized metrics would be more accurate, sorry!).

As such, after profiling both the CPU and IO components, I've picked up the low-hanging fruits to make Heimdall more than just a single-run tool.

Please note, I have added "Mock" in the title because while this works well, it would require some cleaning to be accepted as a legitimate PR. So, see this as some form of "heads-up".

Solution

The solution consisted of three main components:

• Pass around a real cache of function signatures and their respective ResolvedFunction arrays. Indeed, what's defined as cache in the codebase would be better called speculative cache as it is implicitly assumed that once loaded the OS will keep that in memory. Nonetheless, it is good practice not to rely on OS abstraction for this very speculative nature. As such, we load the whole key:value mapping and allow it to be passed around threads.
• Lazy-loading a single HTTP Client that can be passed around threads, since the previous implementation would spin up a client per request. That would result in significant CPU cycles being spent re-doing cryptographic functions of the TLS protocol (notably, SSL Handshake).
• Added optionality to avoid expensive compute such as similarity checks via normalized_damerau_levenshtein. I'd argue that more formal verification is required to understand how much of an improvement these bring versus the 4-5x added latency once the above two are implemented.

Obviously, you'll notice that cache update can be done better, can be flushed at the end of the process into the /cache directory, etc. I am happy with these limitations, as my goal is met here. Ultimately, these elements justify making this PR "mock", and subject to relatively mild, but yet sensible final polishing work.

I've saved some flamegraph and other CPU/IO profiling done along the way (even though I'd argue most of the performance was picked up via flamegraph, profiling on MacOS requires more upfront work than on Linux...). If you want me to pass these around, feel free to ask!

[Jon-Becker]

Hey! Thank you for opening! I'll take a look today :)

If you wouldnt mind sharing profiling results & flamegraphs, you can comment them here or send to jonathan@jbecker.dev!

[sfyll]

Hey Jon!

I'll enclose a flamegraph I made before making changes to Heimdall, and one from the latest iteration (with damerau toggled-off). Please note my code is using both parallelism and concurrency, resolving transactions input in batches of 10,000. You can easily run your own flamegraph using https://github.com/flamegraph-rs/flamegraph, or profile your app using https://github.com/cmyr/cargo-instruments if on MacOS (can't share these as they have sensitive information)

[Jon-Becker]

I've removed the normalized_damerau_levenshtein checks from the decode module on nightly, with the previous fixes & improvements it became obselete and unnecessary!

I'll take a look at implementing more optimizations from this PR shortly <3

[sfyll]

any questions just shoot

thanks a lot for having created heimdall-rs in any cases ! 🫡